ABSTRACT: BACKGROUND: Intimate partner violence (IPV) and termination of pregnancy (TOP) are global health concerns, but their interaction is undetermined. The aim of this study was to determine whether there is an association between IPV and TOP. METHODS AND FINDINGS: A systematic review based on a search of Medline, Embase, PsycINFO, and Ovid Maternity and Infant Care from each database's inception to 21 September 2013 for peer-reviewed articles of any design and language found 74 studies regarding women who had undergone TOP and had experienced at least one domain (physical, sexual, or emotional) of IPV. Prevalence of IPV and association between IPV and TOP were meta-analysed. Sample sizes ranged from eight to 33,385 participants. Worldwide, rates of IPV in the preceding year in women undergoing TOP ranged from 2.5% to 30%. Lifetime prevalence by meta-analysis was shown to be 24.9% (95% CI 19.9% to 30.6%); heterogeneity was high (I (2)>90%), and variation was not explained by study design, quality, or size, or country gross national income per capita. IPV, including history of rape, sexual assault, contraceptive sabotage, and coerced decision-making, was associated with TOP, and with repeat TOPs. By meta-analysis, partner not knowing about the TOP was shown to be significantly associated with IPV (pooled odds ratio 2.97, 95% CI 2.39 to 3.69). Women in violent relationships were more likely to have concealed the TOP from their partner than those who were not. Demographic factors including age, ethnicity, education, marital status, income, employment, and drug and alcohol use showed no strong or consistent mediating effect. Few long-term outcomes were studied. Women welcomed the opportunity to disclose IPV and be offered help. Limitations include study heterogeneity, potential underreporting of both IPV and TOP in primary data sources, and inherent difficulties in validation. CONCLUSIONS: IPV is associated with TOP. Novel public health approaches are required to prevent IPV. TOP services provide an opportune health-based setting to design and test interventions.